Initial clinical trials with LAK cells and IL-2 are associated with partial or complete responses in approximately 20% of patients with renal cell carcinoma and melanoma. However, the comparison of the antitumor activity of IL-2 alone versus IL-2 in combination with LAK cells showed at best only a small clinical benefit from the LAK cells. A recently completed phase I trial at the Frederick CRB using IL-2 and in vitro anti-CD3 activated T cells revealed an unexpectedly low response rate. One limitation of this trial is the fact that only cells that are circulated in the peripheral blood can be activated. Preclinical studies suggest that in a tumor in which CD8+ cells are the mode of rejection, the timing of anti-CD3 and IL-2 injection are of major importance. Based on these considerations, we propose to administer immunologically active doses of anti-CD3 in vivo to cancer pts. simultaneously with bolus and continuous infusion IL-2. The objectives of this study are: to determine whether anti-CD3 can be safely administered in combination with IL-2; to determine the toxicity of this regimen, to measure selective immunomodulatory effect; to observe any antitumor responses, and determine whether expansion of activated T lymphocytes can be achieved in vivo by this method. Pts. will receive anti-CD3 (400, 800, 1200, & 800 mu/m2), continuous infusion IL-2 (IV- IL2 0.75 mu/m2), and bolus IL-2 (1.5 mu/m2). Cohort 4 will receive cytoxan. 12 pts. have been entered on this study to date. There have been 2 PR's. Pts. have sustained significant toxicity including hypotension, chills, fever, and metabolic acidosis. The hypotension has been more pronounced than with similar doses of IL-2 alone. These have been manageable. We plan to accrue a total of 24 pts.